In this tutorial, we have explored how to determine the appropriate sample size for non-inferiority studies. Non-inferiority means testing if a proposed treatment is no worse than an existing approach by showing it is above the non-inferiority bound.
Introduction to Non-inferiority Designs
Important Design Considerations for Non-inferiority Designs
Sample Size Determination for Non-inferiority Designs
Non-inferiority testing is used to test if a new treatment is not inferior to a standard treatment. This is a common objective in areas such as medical devices and generic drug development. For example, if a proposed device or treatment were less invasive than the standard treatment then non-inferiority would be an appropriate route to improve patients’ treatment choices.
To test for non-inferiority, the treatment group is tested to verify it is above the non-inferiority margin. The non-inferiority margin is a level below equality that would still be considered acceptably non-inferior to the standard treatment. The definition of the non-inferiority margin is a matter of significant debate and is an important aspect of regulatory guidance from agencies such as the FDA.
Non-inferiority studies can be conducted for a wide variety of different endpoints including continuous, binomial, survival and count endpoints. Each of these endpoints present unique design and statistical considerations with a wide range of potential design choices. For example, common designs for non-inferiority are crossover designs, three arm trials and parallel arm trials.
In this tutorial, we have reviewed non-inferiority testing design considerations and demonstrated how to determine the sample size for a wide variety of endpoints and designs.